184 ALASKAN GLACIER STUDIES
the stagnant ice forward and breaking it, there could hardly have been so abrupt a termination, for after the first thrust further flowage should cause continued advance. Because of these objections we feel forced to set this hypothesis aside as improbable.
Another conceivable hypothesis is that the sudden addition to the upper portions-of the glaciers started a wave in the ice, which was transmitted down-stream, and which was transmitted wholly independently of glacier motion. As this wave paused down the glacier it broke the ice surface, and since it moved with rapidity it transformed the glacier from end to end in a brief period of time; and after it had passed, the ice returned to its normal state and the broken surface was given over to the processes of ablation to heal. We confess that we find this hypothesis difficult of conception, and that the form and nature of movement of the wave is not clear. If, however, we could grant the passage of such a wave, it would explain some of the phenomena of the advancing Yakutat Bay glaciers, though there are many which are not capable of explanation by such a wave. This hypothesis, for instance, would not explain the notable advance of the front of the glaciers. To carry an ice front forward one mile, as in the case of Haenke Glacier, and two miles, as in the case of the Hidden Glacier, demands an actual transfer of ice. With such a passage of a wave through the glacier, so far as we are able to conceive it, we should expect that when it reached a stagnant piedmont bulb it would spread and gradually die out, as was the case in the Variegated and Atre-vida piedmont areas. But in Malaspina Glacier the breaking extended many miles down the Marvine lobe, without spreading laterally westward; in other words, it followed the flow lines influenced by the Marvine. A wave in the ice might be expected to disregard this factor and pass through ice regardless of its source of origin; and a wave derived from an impulse set up in the Marvine Glacier valley, would, it seems to us, on passing out into the Malaspina Glacier, as easily spread out westward into the ice supplied by the Seward Glacier as southward along the Marvine lobe. A third objection is the thickening of the ice, especially noticeable in the piedmont areas of the Variegated Glacier and the western margin of the Atrevida Glacier. In these places the ice has the appearance of having flowed out there as a thick viscous fluid flows. In this connection may be mentioned the development of the clear ice area in the Atrevida Glacier which we are wholly unable to explain on the hypothesis of the passage of a wave through the glacier. These facts make it evident that a mere wave is inadequate as an explanation of the phenomena of the advancing Yakutat Bay glaciers. Any theory to account for these phenomena must explain a rapid transfer of impulse from far up the glaciers, probably from the reservoir, through many miles of hitherto stagnant or slowly moving ice, and accompanying this a breaking of the glacier surface and an advance of the front, a spreading of the margins and a thickening of the glacier. A combination of the theory of wave motion and actual flowage will account for these phenomena, though neither one will do so by itself, nor, for that matter, will any other theory of which we are cognizant. We are therefore proposing, as an explanation of the behaviour of the Alaskan glaciers an hypothesis which we call the glacier flood hypothesis.
Under the glacier flood hypothesis it is assumed that the ice of a glacier is rigid in its upper portion and along its margins, but that it is in a state of viscosity beneath a rigid crust; that is, a glacier has its zones of fracture and of flowage, like the earth's crust. It is further assumed that the degree of viscosity is increased by pressure, and